Creak Derived from CAPE-V Sentences in Patients with AdLd and pMTD.

OBJECTIVE

The purpose of this study was to evaluate whether acoustic creak (%) derived from the Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V) sentences differentiates adductor laryngeal dystonia (AdLD) from primary muscle tension dysphonia (pMTD).

METHODS

In this retrospective study, acoustic recordings from 42 patients (21 with AdLD and 21 with pMTD) at a tertiary voice center were analyzed. Participants produced six CAPE-V sentences during a standardized voice evaluation. Recordings were segmented and analyzed using an automated algorithm to compute percent creak (%), defined as creak duration relative to total voicing duration. Smoothed cepstral peak prominence (CPP) was also extracted. Hierarchical binomial logistic regression models evaluated whether creak alone or creak in combination with CPP predicted the diagnostic group.

RESULTS

Mean creak (%) values were comparable between groups. Creak (%) did not significantly predict the diagnostic group (P > 0.05). Inclusion of CPP did not improve model performance. Creak distributions were positively skewed in both groups, with notable interindividual variability.

CONCLUSION

Creak derived from CAPE-V sentences does not differentiate AdLD from pMTD. These findings suggest that stimulus characteristics, particularly sentence length and the associated respiratory-phonatory demands, may influence the discriminative validity of creak. Further research is required to clarify its clinical utility across speech tasks.